Skip to main content

Computational prediction and comparative analysis of protein subcellular localization in bacteria

Resource type
Thesis type
(Thesis) Ph.D.
Date created
2006
Authors/Contributors
Abstract
Predicting the subcellular localization of a protein is a critical step in processes ranging from genome annotation to drug and vaccine target discovery. Previously developed methods for localization prediction in bacteria exhibit poor predictive performance and are not conducive to the high-throughput analysis required in this era of genome-scale biological analysis. We therefore developed PSORTb, a hig h-precision, hig h-throug hput tool for the prediction of bacterial protein localization. PSORTb implements a multi-component approach to prediction, incorporating the detection of several sequence features known to influence subcellular localization. With a reported overall precision of 96%, it is the most precise method available and one of the most comprehensive methods capable of assigning a query protein to one or more of four Gram-positive or five Gram-negative localization sites. The PSORTb algorithm comprises a series of analytical steps, each step - or module - being an independent piece of software which scans the protein for the presence or absence of a particular sequence feature. Modules include: SCL-BLAST for homology-based detection, the HMMTOP transmembrane helix prediction tool, a signal peptide prediction tool, a series of frequent subsequence-based support vector machines, as well as motif and profile-matching modules. The modules return as output either a predicted localization site or - if the feature is not detected - a result of unknown'. The output is then integrated by a Bayesian network into a final prediction. Development of PSORTb also required the creation of PSORTdb, a database storing both known and predicted localization information for bacterial proteins. This is a valuable resource to both the localization prediction and microbial research communities, providing a source of training data for new predictive algorithms and acting as a discovery space. The release of PSORTb v.2.0 allowed us to carry out a number of analyses related to localization. We performed the first genome-wide computational and laboratory screen for Nterminal signal peptides in the opportunistic pathogen Pseudomonas aeruginosa, used PSORTb as a complement to laboratory-based high-throughput 2D gel studies of individual cellular compartments, and examined protein localization in a global context, revealing trends with implications for adaptive evolution in microbes.
Document
Copyright statement
Copyright is held by the author.
Permissions
The author has not granted permission for the file to be printed nor for the text to be copied and pasted. If you would like a printable copy of this thesis, please contact summit-permissions@sfu.ca.
Scholarly level
Language
English
Download file Size
etd2263.pdf 1.81 MB

Views & downloads - as of June 2023

Views: 0
Downloads: 0